Thread: The Vibrations - Measured and Weighed

There is something 'strange' about mixing science and lucid dreaming... reminds me of a mix between Flatliners (dying on purpose and being brought back from death) and Dreamscape (1984 movie) ... being able to use a device to trigger lucid dreams would open up interesting possibilities though. Is there something that you are hoping to accomplish with this research ?

My original goal was to accurately detect REM sleep. Now that that is working consistently I have made the data available to other researchers in the hope that new patterns are found that I have missed.

But a Nobel prize would be nice... for science!

You should look over the measurements of ur lucid dreams in particular and tell us how that differs from the normal dream state if at all.

Originally Posted by IAmCoder

This is my first post in beyond dreaming, as I am a man of science. So please go easy on me.

Originally Posted by nina

Not sure why you put this in BD, but moved to Research.

Definitely not going easy on you, man of science!

Did you put this into BD due to a suspected relation between the vibrations you observed and the vibrations described in OBEs as well as WILDing for some people? The only guess I can make so far, otherwise I agree with nina in not knowing why BD.

What I see in the file is interesting, though I don't know exactly how to interpret it. Mind explaining your thought a bit more?

Originally Posted by dreamdimensions

You should look over the measurements of ur lucid dreams in particular and tell us how that differs from the normal dream state if at all.

There is sometimes a pause in the blinking at the onset of lucidity. You are welcome to look over the waves yourself and let me know if you spot anything interesting.

Originally Posted by nina

Not sure why you put this in BD, but moved to Research.

Thanks!

Originally Posted by StaySharp

Definitely not going easy on you, man of science!

Did you put this into BD due to a suspected relation between the vibrations you observed and the vibrations described in OBEs as well as WILDing for some people? The only guess I can make so far, otherwise I agree with nina in not knowing why BD.

What I see in the file is interesting, though I don't know exactly how to interpret it. Mind explaining your thought a bit more?

If you expand the files down to the minute level, you can see the most detail in the data, which is the amount of motion around my head. You can sometimes see the heartbeat quite well and, of course, the blinking during REM. An anthem is played whenever fREM starts.
So, what I have been seeing is that there is a big movement before the blinking starts and sometimes when it ends, for roughly 10 seconds. But it is generally very quite before and after that noise.

And yeah, I always felt that the vibrations were more of a subjective phenomenon - nay, a hallucination. Not something that could be measured and studied scientifically. I guess I should write an article for one of the journals.

What is the fREM (orange) line indicating?

It seems to be very long compared to the others.

Also, a couple of questions just regarding the device in general,
how do you plug it in? Was this the one that needed an Android device?
Or just a computer?

I'm getting very close to buying one

fREM is just when the plugin detects rapid eye movement (when it sees 8 times of FM > 800 for 1-5 ticks, F1 < 600 for 30 to 60 ticks) and also indicates when the anthem was played. I still need to tweak the formula as some dreams slip through (when the blinking is too fast or slow) and make it configurable in the app.

You just plug it into the USB drive. It does not need an Android device. Just a computer or net-book running any version of Windows.

I am getting very close to mass producing them.

That's great that you've almost got enough to mass produce!

One more thing, which one would you say is the best? Based on what I've seen I'm thinking the Pulse Monitor one would be the best indicator of REM?
Hopefully I'm right coz the EEG is twice the price

Thanks man! The best one is the halograph FM. So good, in fact, that I will probably be shelving the others for now.

Oh, I didn't realise the FM was an actual separate device. So how does that one work exactly?

It is an accelerometer wrapped in a headband connected via a USB cable to your PC which runs Lucid Scribe and waits for a blinking pattern to emerge (fREM) before playing a MP3 of your choice. Or your voice.

Oh woops, should've picked that up - Frequency Monitor -

Put in my pre-order

Sweet as!

Moved your pre-order to the top. Less shipping costs and all.

My curiousity in this device is rising a lot as well. How much do you expect it to cost? 256$ would be probably too much for me now but that doesn't really kill my interest. And when do you expect to reach the necessery 100 pre-orders?

Originally Posted by StaySharp

My curiousity in this device is rising a lot as well. How much do you expect it to cost? 256$ would be probably too much for me now but that doesn't really kill my interest. And when do you expect to reach the necessery 100 pre-orders?

I thought $256 is a lot too at first, but... just think about what you'd be spending it on otherwise. Probably something which wouldn't bring any where near as much satisfaction or personal progress. It's about as much as a gaming console, even less when compared to some of them.

I was excited to see LSD and ug too close together Even more exciting is that there's an EILD device that actually WORKS!!! We are entering into an entire new age!

I'm very lucky to find this thread. I've been struggling with a similar issue - detecting REM through actigraphy. Right now I'm working with an iPhone, and got a question about this screenshot. It uses an accelerometer and a gyroscope together to estimate user activity at night.

Paul, from your experience, you mention that there's an equivalent of rolling over at the start and end of REM. I'm looking at the screenshot below, and your discovery sheds light on the mysterious movement late at REM. If I understand your discovery correctly, to estimate REM (the blue Dreaming? blocks), I would be looking for 2 relatively large spikes of movement (around 500 activity counts) or so? How big of a spike should I expect?

PS. Connect this device to an iPhone via bluetooth, and you would be able to retire in a year




PPS. Very sorry if I"m hijacking a thread!

Originally Posted by Ev

I was excited to see LSD and ug too close together

lmao

[/QUOTE]

Just wanted to add this.... Your sleep stages are almost exactly textbook lol

Originally Posted by tommo

It's about as much as a gaming console, even less when compared to some of them.

I like that comparison... then I can say that my graphics are 32 times better than!

Originally Posted by Ev

Even more exciting is that there's an EILD device that actually WORKS!!! We are entering into an entire new age!

Thanks man! I hope they call it the age of light.

Originally Posted by Ev

If I understand your discovery correctly, to estimate REM (the blue Dreaming? blocks), I would be looking for 2 relatively large spikes of movement (around 500 activity counts) or so? How big of a spike should I expect?

No, you would be waiting for a single spike lasting arond 50 seconds. Then you would wait for at least 30 seconds (to avoid waking in sleep paralysis) before triggering. Any spike after that lasting more than a few seconds could signal the end of REM.

You are welcome to download the LSD files and browse to where the dreams are for further study. You can also open them as text files (they are based on XML) to see all the data in there.

I have been meaning to add "Android Lucid Dreaming App" to the hardware category so I can link back to you. And, of course, a few recordings in that category with an extra line from your output.

Any idea what Avi Sadeh, the world expert on Actigraphy, would say now?

Originally Posted by tommo

Just wanted to add this.... Your sleep stages are almost exactly textbook lol

04:38 - 12:58? And I was worried that I was burning too much midnight oil.

You are welcome to download the LSD files and browse to where the dreams are for further study. You can also open them as text files (they are based on XML) to see all the data in there.

I'm not sure as to how to read them. They are a series of graph values sampled at 1/10th of a second, right? Some of the files linked on the site have rows from 0 to 59. This confuses me.

I have been meaning to add "Android Lucid Dreaming App" to the hardware category so I can link back to you. And, of course, a few recordings in that category with an extra line from your output.

So you want to run the lucid dreaming app side by side and compare the outputs?

04:38 - 12:58? And I was worried that I was burning too much midnight oil

Haha, that's UTC timezone, this is actually 0:38 to 8:15 or so in EST.

No, you would be waiting for a single spike lasting arond 50 seconds. Then you would wait for at least 30 seconds (to avoid waking in sleep paralysis) before triggering. Any spike after that lasting more than a few seconds could signal the end of REM.

I want to hug you This is pretty much what I've been looking for! Right now I'm writing an iPhone app, and I reduced the epoch from 60 seconds down to 10 seconds. Within that epoch the data is processed as activity count (x standard deviations above noise). With the old epoch of 60 seconds I would've seen just another spike on the graph at the REM onset on any graph! Looking at the data through epochs, the spikes go up, while it seems what you found is that I should be looking for a long segment of consistently elevated activity.

Now I know how to create an automated method of REM detection. It's something along the lines of 120 second running window which looks for ~50 seconds of elevated activity followed by 30-60 seconds of (should it be inactivity/paralysis?). After that the app would keep switching states according to this rule.

Any idea what Avi Sadeh, the world expert on Actigraphy, would say now?

I dont know, but I'll write to him and link your discovery

Originally Posted by Ev

I'm not sure as to how to read them. They are a series of graph values sampled at 1/10th of a second, right? Some of the files linked on the site have rows from 0 to 59. This confuses me.

I figured only fellow hackers would look that close at the data. Guess I should have made it 1 to 60!

Code:

<Hour N="%Hour(0-23)%"><...><...><M N="%Minute(0-59)%" V="%MinuteValue(0-999)%" />1,2,3,...,597,598,599 // The numbers 1,2,3,...,597,598,599 are the sensor readings for every 100 milliseconds during the minute, and can range from 0 to 999.

Originally Posted by Ev

So you want to run the lucid dreaming app side by side and compare the outputs?

Yes! So I can link back to your site with a lucid dreaming app subcategory. You are also invited to post any interesting recordings to LSDBase.org once I can convert the files for you (and / or your halograph is ready ). I would like to add your account to the contributors group.

Originally Posted by Ev

I want to hug you This is pretty much what I've been looking for!

I am not into that kind of thing. I am sorry if I have given you the wrong impression.

Originally Posted by Ev

Now I know how to create an automated method of REM detection. It's something along the lines of 120 second running window which looks for ~50 seconds of elevated activity followed by 30-60 seconds of (should it be inactivity/paralysis?).

Awesome! Although vibrations detection may be a more accurate description, but your rule is looking very solid now. I believe the paralysis lasts for the entire REM period / runs parallel to it, until another brief period of a few seconds of elevated activity.

Originally Posted by Ev

I dont know, but I'll write to him and link your discovery

Sweet, let us know what he writes back!

I've opened the .LSD files in the Lucid Scribe, and they are in an interesting format. This is XML with hour tags with nested minute tags, each minute has a number attribute and [probably number of values] attribute. Then there's a CDATA block with comma delimited values.

I don't know of an easy way to open this or convert this into a .csv file to look at the data. Are you by any chance familiar with JSON? I find that format to be a lot easier to work with. Using GSON library, I can convert entire Java Value Objects into plain text.
The best thing is that there's no need to write a parser to get the data.


Is there a way to look at multiple minutes at once in Lucid Scribe? I checked the interesting points that you talked about on your blog, but some of them span multiple minutes, so it's hard for me to get a good grasp of the accelerometer dynamics for those events.

There currently is not a way to look at multiple minutes in Lucid Scribe. I will add an option to the hour view to do this in the next release.

For now, you could in theory copy the minute data from all the minutes you are interested in and append them to one minute inside the LSD file. When you then select the minute that you have pasted everything into, it should plot a span of multiple minutes (after saving the file).